Download Modelling the impact of climate change and weather related events

Modelling the impact of climate
change and weather related events
Seong Woh Choo
Head of R&D and
Chief Underwriting
Officer
1
1
Introduction
•
Risk to community and insurers
•
Climatic cycles/change affecting Australia
•
Weather Risk projects
•
Business initiatives
2
Main ideas
•
Weather Risk Modelling
–
Adds value
–
Assists business and community
–
Can make a difference
3
Insurance Australia Group Limited
• Largest General Insurer in Australia and NZ
• GWP in excess US$3B
• High market share in personal lines
4
IAG’s R&D capability
• 80 analysts with diverse expertise
• Work with external groups
• Used SAS® for over 20 years
5
Risk to community and insurers
• Worldwide 1960-1999
– 85% catastrophes were weather related
– 75% of economic losses
– 87% insured losses
• Losses typically caused by:
– Tropical cyclones/hurricanes
– Winter storms
– Hailstorms
– Bushfires
– Floods
6
H
ai
Insured Losses in $Millions (2001 prices)
ls
to
Ea rm
rt
hq Sy
ua dn
ke ey
H
N Ap
ai
ew ri
ls
C
to
yc ca l 19
rm lo
s
9
ne tle 9
S
B
yd
Tr 19
us
n
ac 89
hf
C
e
ire yc y M y 1
l
H s A on arc 974
ai
e
ls sh
W h1
to
W
an 99
r
e
m
H
dn da 0
ai
B
1
e
ls
r
to isb sda 974
H
ai rm ane y 1
ls
Sy
J 98
to
rm dn an 3
e
Sy y J 198
5
a
d
C
yc ne n 1
y
lo
O 991
n
C
e
c
M t1
Fl y c
9
lo a
o
H od ne dge 86
ai
1
ls s S A l
to
yd th 973
ea
H
a i rm n e
19
y
l
S
s
H
ai tor yd No 71
n
ls
v
to m S ey
19
rm y
dn Nov 84
A
ey
r
19
B mid
F
7
us
al eb 6
hf
e
ire Se 199
pt 2
s
H
ob 19
ar 96
t1
96
7
Australian Catastrophe losses
1800
1600
1400
1200
1000
800
600
400
200
0
7
8
No: 1 - Sydney Hailstorm 1999
• Australia’s most costly natural disaster
• AU$1.7 Billion in Insured Losses
• Contributed 25% of IAG’s 15-year total of weather
related claims
• Approx 50,000 claims
• 4 times bigger than previous largest hail event
9
Sydney hailstorm 1999: 8cm hail
10
No:2- Cyclone Tracy: Darwin 1974
65 killed
600 injured
217 km/hr+ winds
11
Climate Cycles: El-Nino/La Nina
12
Climate Cycles
13
Impact of El-Nino in Australia
14
Impact of La-Nina in Australia
15
Weather Cycles
16
Climate Change: Global
SPM 1a
Temperatures have increased by 0.6°C last century with the 1990’s the warmest decade and 1998 the warmest year since
1861(instrumental record) A further 1.4 – 5.8 °C increase projected by 2100
17
Climate Change: Australia
18
Climate Change
• Increases in temperature
• more intense rain and wind storms
• increase in coastal flooding from storm surge
• more intense wet and dry spells
• Climate change is expected to increase BOTH
the frequency and severity of claims incurred
19
Sensitivities
Hazard
Cause Of Change In
Hazard
Resulting Change In
Damage/Loss
Windstorm
Doubling of windspeed
2.2 °C mean temperature
increase
Four-fold increase in
damages
Increase of 5-10% in
hurricane wind speeds
Extreme
temperature
episodes
1 °C mean temperature
increase
300-year temperature events
occur every 10 years
Floods
25% increase in 30
minute precipitation
Flooding return period
reduced from 100 years to 17
years
Bushfire
1 °C mean summer
temperature increase
Doubling of CO2
17-28% increase wildfires
143% increase in
catastrophic wildfires
20
Impact of Peak Gust Changes
700
% Increase in Damages
600
500
25% increase in peak gust causes
650% increase in building damages
400
300
200
100
0
Under 20 knots
20-40 knots
40-50 knots
50-60 knots
IAG Building claims versus peak gust speed showing disproportionate increase in claims cost from small increases in peak gust speed.
21
Data and Analysis Tools
Data Extraction
Data Management
Data Exploitation
based Data
Daily Extracts
SAS®
Data
Warehouse
Data Mining
Enterprise Miner®
Analysis data
DB2 Transaction
Daily Extraction, Transformation & Loading
(Using Defined Business Rules)
Reporting/
Summarisation
Base SAS®
External
Climate
Data
Statistical and Data
Analysis
SAS/STAT®
SAS/insight®
Matlab®
Presentation/
Interpretation
Base SAS®
Enterprise Miner®
Mapinfo®
Microsoft Excel
22
IAG’s Storm Modelling
• Larger or more intense ?
• Characteristics and behavior change?
• More likely ?
23
Hail/Extreme Events Modelling
24
Could climate change yield a “mega-storm”?
Possible
Storm with
8cm Hail?
8cm footprint
for the April
1999
Storm
25
How could this happen?
•
Produced by small changes to three key parameters
– Wind-shear
– Sea surface temperature
– Atmospheric stability
Related to climate Change?
•
Sea surface temperature:
Yes
Wind Shear
:
Probably
Atmospheric stability
:
Probably
26
Modelling collision frequencies
27
Business response
•
Recognition climate change is a global phenomena
• Support initiatives that reduce greenhouse
emissions
•
Models add value
• helps business understand the impact
• Research continuing
Working with Govt and community
- Affordability
- Building codes
28
Business response
•
Example of initiatives
• Education program on sustainability
• Early warning broadcast on storm
• Recycling program with repairers
• 2004 target – reduction of
• 15% on paper used
• 5% in fuel consumption
• 5% in kms of air travel
29
Summary
Weather risk modelling is feasible
Adds significant value
Assists business and community
You can make a difference
30
Thank you
Questions?
31